Most Important Update
Alibaba has released Qwen3-Coder-Next, a new open-weight AI model for programming agents and local development. Trained on 800,000 verifiable tasks, the model has 80 billion parameters total but only 3 billion active at any time. It outperforms or matches much larger open-source models on coding benchmarks, scoring above 70 percent on SWE-Bench Verified with the SWE-Agent framework. [1]
Coding-Focused Updates
- Alibaba's Qwen3-Coder-Next is available on Hugging Face and ModelScope under the Apache 2.0 license.
- Qwen3-Coder-Next supports 256,000 tokens of context and works with development environments like Claude Code, Qwen Code, Qoder, Kilo, Trae, and Cline.
- Local tools like Ollama, LMStudio, MLX-LM, llama.cpp, and KTransformers also support Qwen3-Coder-Next.
New Model Releases
- Qwen3-Coder-Next, an open-weight language model designed specifically for coding agents and local development, has been released by Alibaba.
- Qwen3-Coder-Next is built on top of Qwen3-Next-80B-A3B-Base, which adopts a novel architecture with hybrid attention and MoE.
General AI Updates
- There are no significant general AI updates to report.
Model Comparison Table
| Model | Parameters | Active Parameters | Benchmarks |
|---|---|---|---|
| Qwen3-Coder-Next | 80 billion | 3 billion | SWE-Bench Verified: 70%+ |
Best News for Developers
Qwen3-Coder-Next is a compact AI model that outperforms or matches much larger open-source models on coding benchmarks, making it a valuable tool for developers.
Best News for Students
Qwen3-Coder-Next is an open-weight AI model that is available on Hugging Face and ModelScope, making it accessible for students to use and learn from.
Rumors and Weak Claims
None.